Clariverdict/Index/Open Source AI Models/Mixtral 8x22B

Mixtral 8x22B

FreeOpen Source AI Models

Category:Open Source AI Models·Pricing:Free

Last Updated:N/ALast Verified:Aug 14, 2026Source:View source

Mistral's Mixture-of-Experts (MoE) open-weights powerhouse with exceptional throughput.

Freshness 50%Aug 14, 20261 source

#Mistral AI#Mixture of Experts#High Throughput#Math Core

Visit Website

Overview

Mixtral 8x22B is a sparse Mixture-of-Experts model. It activates only a fraction of its weights per token, achieving high reasoning speeds and deep mathematical precision.

Features

Sparse Mixture-of-Experts (MoE) routing engine

Active context buffer reaching 64,000 tokens

Strong coding, mathematics, and translation benchmarks

Fully commercially permissive Apache 2.0 license weight files

Editorial Transparency: Every link on Clariverdict directs to the official landing page of the tool. We do not host paid search placements, pay-to-play ranks, or modified tracking URLs.

Details

Primary Categoryopen-source

Billing TypeFree

Updates Logs0 releases

Verification

Official sourceLinked

Last verifiedAug 14, 2026

StatusVerified

Frequently asked questions

Is Mixtral 8x22B free?

Pricing model: Free. Free weight downloads under Apache 2.0 licensing. Cloud hosting options available.

What is Mixtral 8x22B best for?

Mistral's Mixture-of-Experts (MoE) open-weights powerhouse with exceptional throughput.

What are alternatives to Mixtral 8x22B?

Documented alternatives in the Clariverdict catalog: Mistral, Llama 3.1 405B. Open each profile for verified specs, pricing, and sources.

When was Mixtral 8x22B last verified?

Clariverdict last verified this listing on 2026-08-14 against the official source at https://mistral.ai.